Postdocs

While other fields have standardized postdoc programs, computer science still searches for the right approach for postdocs. While we have had postdoc positions since I can remember, quite often researchers have gone straight from Ph.D. to tenure-track positions particularly in times of high growth in CS departments (early-mid 80s and mid-late 90s).

We are now seeing a spike in the demand for postdoc positions for several reasons.

• A tightening job market means less tenure-track jobs so more people opt to do postdocs to build up their CVs. Several researchers are even taking second and third postdocs, not long ago a rarity in CS.
• Many students opt to delay a tenure-track position for a year and do a postdoc first. The commitment goes both ways, if a department is holding a position for a student then that student is committing to going to that department. It's not fair to go back on the job market during that postdoc year. Some departments are becoming more reluctant to allow the year delay because of bad experiences with students not fulfilling that commitment.
• More and more students attend graduate school in their home countries and hope to eventually return to permanent jobs in those countries but take postdoc positions elsewhere to get a broader view of the field.

Alas we don't have an increase in postdoc supply to go along with this demand. Many of the industrial research labs have shrunk and hire few or no postdocs. Meanwhile most US academic departments have no permanent postdoc programs and CS grants are rarely large enough to cover the cost of a postdoc (as opposed to Canadian and European groups which tend to have more postdocs). Just another way the US is losing strong researchers to other countries.